View Single Post
  #12   Report Post  
jim
 
Posts: n/a
Default


"Ian Iveson" wrote in message
...
"jim" wrote

I've seen this in done in a few '60s schematics and 'Glass Audio'

articles.
How do you get the ultimate out of as few valves as possible ??
Global NFB reduces distortion but decreases gain.. Add PFB at the

OP valve
drivers. You get the gain back and the global NFB keeps the

distortion
low...

You balance one against the other...

Theory says it works. Reality says it's unstable and

unpredictable.


The quad 2 uses local +ve fb to one half of the floating paraphase
splitter, and -ve to the other half, and -ve overall. It's British.
Must be a good thing.

cheers, Ian



The Austin Allegro was British too ! Hi, Ian....
I have a reprint of a Glass Audio article somewhere headed 'A balanced
feedback amplifier'
PP EL84s if I remember right. A Williamson .The author achieved high
sensitivity by making the PFB adjustable using a preset
pot, winding it up until the amp oscillated and then backing it off. Some
impressive claims were made for the circuit but no details of the test
procedures or the effect of different output loads were mentioned
Analysis of the paraphase is somewhat difficult (particularly for me ! ) It
seems to get drawn in a number of different ways, as well. Taking a
proportion of the output of V1 to drive V2 with an opposite phase signal
using a resistance divider proportioned to be equal to the gain of V1 is
pretty obvious but when it becomes floating and self balancing, these
unequal resistances disappear. Using equal resistances, V2 grid is driven
by the difference between the two outputs, which appears across the common
resistor which ties V2 grid down to earth. Balance should therefore be as
accurate as the matching of the two resistors. Differences in the
characteristics of the two triodes should theoretically be cancelled out
too. HT is not wasted, as it is across the tail resistor of an LTP, and the
paraphase should therefore be capable of producing larger output signals.
Remember, I know bugger all, so such an analysis is probably completely
wrong
kind regards
jim